[Bug]: The liquid will not leak from the bottom of the container when it is placed close to the ground, but it will seep out in a container that is suspended in the air.

[Kawhi5]

Bug Description

When the height parameter of the container is 0, the liquid inside will not leak. However, as the height of the container is increased, the liquid will leak from the bottom.

86a7970368fef027e7d826f6a3a9ba52_raw.mp4

Expected Behavior

Provide some assistance and references regarding liquid leakage.

[duburcqa]

@Kawhi5 Please could you share a standalone reproduction script?

[Kawhi5]

@Kawhi5 Please could you share a standalone reproduction script?
Here are relevant materials. Thank you very much for your reply!

125test.py
illustration.txt
liquid.py
WindJet.py

[duburcqa]

Some files are missing!

    file  = 'D:\Genesis-main-20250828A\genesis/assets\meshes/new deep hole.obj',
    file  = 'D:\Genesis-main-20250828A\genesis/assets\meshes/KN-R02-150.obj',

[Kawhi5]

Some files are missing!

    file  = 'D:\Genesis-main-20250828A\genesis/assets\meshes/new deep hole.obj',
    file  = 'D:\Genesis-main-20250828A\genesis/assets\meshes/KN-R02-150.obj',

Sorry, sir. GitHub does not support OBJ format files. Could you please provide me with your email address?

[duburcqa]

You should be able to compress them as a zip, tar or tar archive.

[Kawhi5]

You should be able to compress them as a zip, tar or tar archive.

script.zip
I'm sorry to trouble you.

[hughperkins]

Hi Kawhi5,

1. for the height issue, this affect is caused by the existence of a plane a y=0
   • when the container is at height=0, the plane covers the bottom end of the container, and stops the fluid leaking out
2. as for why the fluid leaks out, I made two changes, which together gave me a fluid that stayed in the container reasonably well I feel:
   • increased the scale by 10x, and
   • used a much simpler mesh

(Note that I removed the plane, and the wind generator, because, with the plane, it hides the leak; and even without the wind generator, the container was leaking for me, so I felt let's keep things simple for now :) )

For the first change, here are screenshots of your scene scaled up by 10x:

(Corresponding script: 2319/script/125test.py, in attached zip file)

• we can see that the container holds the liquid better, but there's still a strange 'ballooning' effect, and liquid leaks out the bottom, in a narrow stream

For the second change, combined with scaling up, here are screenshots of the results using a very simple cubic mesh:

(Corresponding script: 2319/test_box_container.py , in attached zip file)

Wire-frame comparison between the two meshes:

So, my hypotheses:

• the existence of the plane at y=0 is the reason for the perceived 'height' effect
• the scale is currently too small for mpm liquids to work well (though you could also try reducing the timestep size, if you really must use small dimensions, perhaps)
• the existing mesh is complex, and might be hard for the physics simulator to work well with

Thoughts?

Hugh

zipfile:

2319.zip

[Kawhi5]

Hi Kawhi5,

1. for the height issue, this affect is caused by the existence of a plane a y=0

   • when the container is at height=0, the plane covers the bottom end of the container, and stops the fluid leaking out

2. as for why the fluid leaks out, I made two changes, which together gave me a fluid that stayed in the container reasonably well I feel:

   • increased the scale by 10x, and
   • used a much simpler mesh

(Note that I removed the plane, and the wind generator, because, with the plane, it hides the leak; and even without the wind generator, the container was leaking for me, so I felt let's keep things simple for now :) )

For the first change, here are screenshots of your scene scaled up by 10x:

(Corresponding script: 2319/script/125test.py, in attached zip file)

• we can see that the container holds the liquid better, but there's still a strange 'ballooning' effect, and liquid leaks out the bottom, in a narrow stream

For the second change, combined with scaling up, here are screenshots of the results using a very simple cubic mesh:

(Corresponding script: 2319/test_box_container.py , in attached zip file)

Wire-frame comparison between the two meshes:

So, my hypotheses:

• the existence of the plane at y=0 is the reason for the perceived 'height' effect
• the scale is currently too small for mpm liquids to work well (though you could also try reducing the timestep size, if you really must use small dimensions, perhaps)
• the existing mesh is complex, and might be hard for the physics simulator to work well with

Thoughts?

Hugh

zipfile:

2319.zip

Dear Sir, thank you very much for your reply. I'm sorry to bother you again. The method you suggested has helped solve some problems. However, at present, I need to keep the size ratio of the model unchanged (50x50x100mm), so it cannot be enlarged. Secondly, I regard it as a cylindrical water tank. The material of the water tank is similar to glass or plastic. When adding the wind field, it is necessary to ensure that the bottom of the water tank does not leak or expand, and the water particles should flow out through the tank opening. If you have any other suggestions, please feel free to contact me at any time. Thank you ☺️ (Email: wang_pinye@163.com)